Preferential depletion of corrinoids on transcobalamin II (i.e., sharply reduced holo transcobalamin II (TC II)) occurs early in vitamin B12 deficiency. We measured corrinoids (Cor) and cobalamins (Cbl) on transcobalamins I and III (TC I + III) and on TC II. We also measured the unsaturated B12 binding capacities of transcobalamin I and III and TC II in serum from patients with B12 deficiency (N = 5) (with or without concurrent folate deficiency), with pernicious anemia in remission (N = 7) (1 month after therapy), and in several control groups including healthy volunteers (N = 6), hematologically normal elderly hospitalized patients (N = 5), and non-B12 nonfolate deficient anemic elderly hospitalized volunteers (N = 5). In B12 deficient patients, Cor = 177 +/- 92 pg/ml, Cbl = 56 +/- 20 pg/ml, TC II Cor = 1.0 +/- 2.2 pg/ml, and TC II Cbl = 4.4 +/- 4.9 pg/ml in contrast to pooled controls with Cor = 730 +/- 229, Cbl = 523 +/- 198, TC II Cor = 100 +/- 84, and TC II Cbl = 88 +/- 70 (all values expressed in picograms/milliliters). In pernicious anemia in remission, Cor = 505 +/- 138, Cbl = 294 +/- 77, TC II Cor = 80 +/- 31 and TC II Cbl = 37 +/- 36. TC II unsaturated B12 binding capacity was significantly higher in B12 deficient patients than in pooled controls. These data support that: (a) holo TC II is sharply depleted in untreated B12 deficiency; (b) normally, the only Cor on TC II are cobalamins; (c) in treated pernicious anemia, TC II appears to also bind non-cobalamin corrinoids; (d) continued malabsorption of vitamin B12 may result in reduced B12 on TC II within a month after the last parenteral therapy with 1000 micrograms of cyanocobalamin, and (e) TC II UBBC rises as B12 deficiency is developing. Further investigation is required for definitive delineation of whether sharply reduced Cor on TC II in untreated B12 deficiency can diagnose "true" B12 deficiency, in view of false positive or false negative results which occur in all serum B12 assays.